Bridging the gap: explainable ai for autism diagnosis and parental support with TabPFNMix and SHAP —an article seemingly written with AI assistance and then presumably reviewed with AI assistance–was published by Scientific Reports (from Nature Group) on November 19th, 2025. It took nine days for the editor to update the article with a note about the curious content. The centerpiece of the article was Figure 1, featuring an improbable bicycle seat, an equally improbable leg hole table seat, and the exhortation Factor Fexcectorn!
This article has generated a lot of interest at PubPeer, the online journal club, where contributors have noted that in addition to the bizarre figure, the article is based on a model that is not well described, is formatted as if written by a chatbot with bulleted text, and includes many citations that do not support the referenced sections (and some citations that may not exist!).
The journal claims to have “rigorous policies to govern every stage and aspect of our peer review process, from initial submission through to revisions, decisions and appeals” and sends papers to external reviewers once they have been initially screened by a member of the editorial board, per their peer-review policies. A response from the editor regarding this particular article is provided at nobreakthroughs,a Substack curated by a former CNET Science Editor: “Whilst the details of peer review are confidential, we can confirm that the article underwent two rounds of review from two independent peer reviewers, supporting an accept decision. We have also done an assessment of the handling of other papers with the same handling editor and are confident that they have a robust record and that this was a case of human error.”
The current article processing charge or charge to the author for the privilege of publishing for an article in Scientific Reports is $2690. You can read more about open access publishing at Emory here: https://libraries.emory.edu/research/open-access-publishing/emory-oa-fund
Naturally, this article has generated a lot of buzz on Altmetrics and while it bears repeating that there’s no such thing as bad publicity, in this case maybe not so much.
This is a good reminder to examine all evidence critically, even if it’s published by a renowned journal.